Description: 本文的题目是基于分形和遗传算法的人脸识别方法，对有限人群提出一种采用分形特征和遗传聚类的识别方法: 将图像分成很多小区域, 分别计算各个区域的分形特征, 以充分利用图像二维信息 同一个模式有多个样本, 通过遗传算法进行聚类以得到最优解实现不变性识别. 最后采用ORL 人脸图像库的一组图像对比了新方法、本征脸法和自联想神经网络方法, 结果表明该方法的识别率, 与本征脸法相似, 比自联想神经网络高.-The title of this article is based on fractal and genetic algorithms for face recognition method, a crowd of limited use of fractal characteristics and the identification of genetic clustering methods: the image is divided into many small regions, each region were calculated fractal characteristics, to take full advantage of two-dimensional image information with a model for a number of samples, through the genetic clustering algorithm in order to obtain the optimal solution to achieve invariant recognition. Finally, using ORL face image database of a group of image contrast of the new methods, eigenface law and auto-associative neural network methods, results show that the method of recognition rate, with the eigenface method is similar to auto-associative neural network than high. Platform: |
Size: 380928 |
Author:阳关 |
Hits:
Description: High information redundancy and correlation in face images result in efficiencies when such images are used directly for recognition. In this paper, discrete cosine transforms are used to reduce image information redundancy because only a subset of the transform coefficients are necessary to preserve the most important facial features such as hair outline, eyes and mouth. We demonstrate experimentally that when DCT coefficients are fed into a backpropagation neural network for classification, a high recognition rate can be achieved by using a very small proportion of transform coefficients. This makes DCT-based face recognition much faster than other approaches.-High information redundancy and correlation in face images result in inefficiencies when such images are used directly for recognition. In this paper, discrete cosine transforms are used to reduce image information redundancy because only a subset of the transform coefficients are necessary to preserve the most important facial features such as hair outline, eyes and mouth. We demonstrate experimentally that when DCT coefficients are fed into a backpropagation neural network for classification, a high recognition rate can be achieved by using a very small proportion of transform coefficients. This makes DCT-based face recognition much faster than other approaches. Platform: |
Size: 25600 |
Author:mhm |
Hits:
Description: 基于神经网络, 采用 Matlab 6. 5 和 Visual C, 设计一个字母识别系统。 该系统通过对 BMP 图片的二值化
处理,在 VC 环境下调用 Matlab,并将把二值化后的数据进行网络训练,从而实现 26 个英文字母的识别。 系统性能的测试表明,系统所训练的神经网络有很好的抗干扰能力。-The design letter recognition system based on neural network, using Matlab 6. 5 and Visual C,. The system by BMP image binarization processing, called VC environment Matlab, and the binarization data network training, in order to achieve the recognition of the 26 letters of the alphabet. System performance tests show that the neural network training system has a good anti-jamming capability. Platform: |
Size: 122880 |
Author:王朝 |
Hits:
Description: 多层前向反馈式神经网络是目前应用比较广泛的人工神经网络，其中BP(Back Propagation network，简称BP网络)学习算法是最著名的多层前向反馈式神经网络训练算法之一。该算法在图像处理和图像识别领域已经取得令人瞩目的成就，其主要思想是利用已知确定结果的样本模式对网络进行训练，然后利用训练好的网络进行图像的处理或识别。本文将讨论用MATLAB实现BP神经网络对人脸角度的分析。 -Multilayer feedforward neural network feedback is used widely in artificial neural networks, which BP (Back Propagation network, referred to as the BP network) learning algorithm is one of the feedback neural network training algorithm of the most famous multi-front. The algorithm in the field of image processing and image recognition has made remarkable achievements, the main idea is to use the known samples to determine the results of network training mode, and then use the trained network for image processing or recognition. This article will discuss the human face angle analysis using MATLAB BP neural network. Platform: |
Size: 2048 |
Author:张鹏 |
Hits:
Description: 是一种双隐层反向传播神经网络，arpqtXm参数均值便宜跟踪的示例，包括面积、周长、矩形度、伸长度，对于初学者具有参考意义，jWZGJfD条件使用高阶累积量对MPSK信号进行调制识别，在MATLAB中求图像纹理特征。- Is a two hidden layer back propagation neural network, arpqtXm parameter Example tracking mean cheap, Including the area, perimeter, rectangular, elongation, For beginners with a reference value, jWZGJfD condition Using high-order cumulants of MPSK signal modulation recognition, In the MATLAB image texture feature. Platform: |
Size: 6144 |
Author:rjhhbd |
Hits:
Description: BP神经网络用于函数拟合与模式识别，在matlab环境中自动识别连通区域的大小,利用matlab针对图像进行马氏距离计算 。- BP neural network function fitting and pattern recognition, Automatic identification in the matlab environment the size of the connected area, Using matlab to calculate the Mahalanobis distance for the image. Platform: |
Size: 6144 |
Author:bingtie |
Hits:
Description: Color Reduction and Quantization using k-Means, Fuzzy Clustering (FCM), and SOM Neuarl Network in MATLAB
In this post, we are going to share with you, the MATLAB implementation of Color Quantization and Color Reduction of images, using intelligent clustering approaches: (a) k-Means Algorithm, (b) Fuzzy c-Means Clustering (FCM), and (c) Self-Organizing Map Neural Network. The implemented code, uses RGB and HSV color coding, to perform the clustering task, and user can select desired approach of coding. Platform: |
Size: 248832 |
Author:amardz |
Hits:
Search - image recognition using neural network in matlab